Chapter 2

Objectivity

measurements needs to be independent.

Measure of entity or behaviour that within an allowed margin of error, is consistant across instruments or observers.

variable

refers to object, concept, or event being measured ex - behaviour, blood/saliva, self reporting, MRI

operational definitions

statements that describe the procedures (or operations) or specific measures/ methods that are used to record observations.

ex - intoxication (variable) - definition (either physiological measure/ blood, behaviour/ walking in live or self reported measure/score on scale)

Reliability - 3 methods

is when a measure provides consistent and stable answers across multiple observations or points in test.

1. test-retest reliability (test),

2. Alternate forms of reliability (alternate version of test that studies same thing)

3. Inter-rater reliability (multiple observers of a behaviour should all come to the same conclusion)

validity

is the degree to which an instrument or procedure actually measures what it claims to measure. ex - intelligence test to test intelligence

generalizability

refers to degree to which one set of results can be applied to other situations. individuals or events. primary way to study this is large study groups because its a sample to population and easy to generalize.

Critical evaluation of findings, beware of cases of over-generalization - 'mozart effect' (people believe that if u listen to classical music makes you more intelligent - found that people did better on spacial tests after listening).

effects of generalizability of results

1. sample type 2. location of study

sample type - 2 types

1. random sample - every individual of population has an equal change of being included.

2. 2. convince sample - samples of individuals who are readily available.

location of study

lab (control, create precise questions) vs natural research (less control, people behaves in natural way).

Ecological validity - the degree to which the results of a lab study can be applied or repeated in the natural environment

Sources of bias in research

1. research bias (to encourage one outcome over others)

2. 2. subject bias (werent behaving naturally)

3. . ex - hawthorne effect - used to describe situations in which behaviour changes as a result of being observed, when they changed certain aspects of working and saw progress increase but realized that when they noticed someone watching them, they got work done.

Demand characteristics

unintentional clues particiants can use to figure out what the study is about, then they aren't behaving in a natural way and invalidating results.

ex - Clever Hans effect - claimed that horse could do crazy shit so to test the validity, he blinded him and varied the people who didn't know the answers. all of these were clues for the horse

social desirability responding

research participants respond in ways that increase the chances they will be viewed favourably.

Can minimize through assurances of anonymous / confidential questioning.

Observer Expectancy effect

researchers expectations can influence subjects behaviour. ex - teacher favouritism

Placebo effect

a measurable and experienced improvement in health or behaviour despite not getting any medication or treatment.

Some evidence shows there is a physiological pain relief, in other cases its completely subjective.

If we have an expectation that this pill is gonna help us, it relieves us which can lead to feeling better. This also works as conditioning effects.

Nocebo effect - become preoccupied with side effects of a drug often become more likely to experience them

things that reduce bias - ACIpSbBb

1. Anonymity

2. 2. Confidentiality 3

3. . Inform Participants. 4.

4. single blind study.

5. . double blind study

anonymity

each individuals responses are recorded without any name or other personal information that could link a individual to specific results

Confidentiality

means the results are only seen by researcher

Inform participants

inform about how the data will be used can reduce anxiety and social desirability bias.

single blind study

participants do not know the true purpose of the study, or else do not know which type of treatment they are receiving (placebo or treatment drug)

double blind study

a study in which neither the participant nor the researcher know the exact treatment for any individual.

peer review

process which papers submitted for publication in scholarly journals are read and critiqued by experts in the specific field of study and to either accept or reject the review.

replication

Once the study is published, the process of repeating a study and finding a similar outcome each time.

the replication crisis

chance results can and do happen.

publication bias results in studies with affirmative evidence being more likely to be accepted into academic journals.

What to believe? the original report or the failed replication. Studies can fail to be replicated for any number of reasons. We must base our beliefs on the entirety of the evidence.

weak forms of evidence

1. anecdotal evidence - an individuals story about an observation or event that is used to make a claim as evidence.

2. ‘2. appeal to authority - the beliefs in an 'expert' claim even when no supporting data or scientific evidence is present ( biased expert).

3. 3. appeal to common sense - tradition or novelty.

4. 4. selective use of data - stats are often inappropriately used to bolster weak arguments.

descriptive data

from observations, no attempts to explain the why. Generated from -

1. case studies

2. naturalistic observations

3. Surveys and questionnaires

case studies

in depth reports about details of a specific case. difficult to generalize findings. ex - phineas gage - drastic personality changes

naturalistic observation

when psychologists unobtrusively observe and record behaviour as it occurs in the subjects natural environment. ex - monkey researcher looking at behaviour

survey and questionnaires

participants makes the observations

correlational research

involves measuring the degree association between two variables.

can calculate the correlation coefficient - direction (positive, going same direction or negative) and magnitude (between range of -1 to 1+, closer to either extreme means strong correlation, if near 0, then a weak correlation). CORRELATION DOES NOT EQUAL CAUSATION, there could be a third hidden variable.

Illusory Correlations

relationships that really exists only in the mind rather in reality. ex - crime increases when the moon is full, opposites attract,

Experimental research - 2 types

variable a causes variable b.

Random assignment - divides samples into two or more groups (experimental and control).

Experimental control - variables are manipulated.

Confounding variables

variables outsides of the researchers control that might effect the results

Between subjects design

participants who are in different groups are compared. a large sample and random assignment makes equal groups likely, not guaranteed.

within subjects design

all participants respond to all types of stimuli or experience all experimental conditions. dont have to worry about inherent differences between peeople as people are acting as control.

cons - order effects (doing one condition before the other can contaminate the other),

• methods to counter act - separating measurements in time,

• counterbalancing (have two groups, one starts with the experiment then moves to control and the other starts with control and ends with experiment and then to compare results).

quasi - experimental research

used when studying groups that have pre determined categories built in, rather than random assignment. ex - comparing men and women. con - cannot determine cause and effect

Research ethics board (REB)

a committee of researchers and officials at an institution charged with the protection of human research participants.

Weighing risks and benefits

1. potential stress to participants (physical, cognitive and emotional stress, to write about traumatic experiences).

2. 2. stress is usually minor, benefits need to outweigh the risks

informed consent

a potential volunteer must be informed of the experiment, and its purpose, tasks, and risks involved in the study. and give consent to participate based on the information provided ex - topic, nature of stimuli, tasks, duration, risks, steps taken to minimize risks

deception

misleading or only partially informing participants of the true topic or hypothesis under investigation (given enough information to consent).

full consent - can refuse participation without fear or penalty, given equal opportunities, given the right to withdrawal, withhold responses.

Debriefing

means that the researchers should explain the true nature of the study, and especially the nature of and reason for the deception

welfare of animals in research

animals are used when -

1. treatments cannot be applied to humans for ethical reasons.

2. Heritability studies require species with short lifespans,

3. examining evolutionary origins of behaviour and cognition.

Committees oversee ethical treatment - given appropriate housing, feeding and sanitation, studies of sick or stressed animals do not provide generalizable results. Risk discomfort needs to be justified and managed humanely

ethical collection, storage and reporting of data

data kept for 3-5 years for any other researcher to repeat the study. TO be honest with the data, acknowledge conflicts of interest.

scientific misconduct has lasting repercussions - the unfounded panic surrounding vaccines and autism.

descriptive stats

a set of techniques used to organize, summarize, and interpret the data.

1. frequency 2. central tendency 3. variability

frequency

the number of observations that fall within a certain category or range of scores.

1. normal distribution (symmetrical)

2. negatively skewed distribution (higher point to the right)

3. positive skewed distribution (higher point to the left)

central tendency

a measure of the central most point of distribution (mean, median, and mode)

variability

the degree to which scores are dispersed in a distribution (high peak, less variability on graph).

standard deviation -> a measure of variability around the mean.

hypothesis testing

a statistical method of evaluating whether differences among groups are meaningful, or could have been arrived at by chance score.

limits - the more tests makes it more likely that you will find a difference that is a fluke and the more sensitive to find significant results (it does not mean its a practical significance) or the acceptable p value is arbitrary.

statistical significance

implies that the mean of the groups are farther apart than you would expect them to be random chance alone.

effect sizes

more nuanced aclternative to significants, alculated value indicates the degree of the difference between groups rather than reducing the significance decision to a yes / no decision